Filtered delay-subspace approach for pilot assisted channel estimation in OFDM systems

Abstract

In this paper we propose a pilot-based OFDM channel estimator based on the combination of low-pass filtering and delay-subspace projection. The proposed estimator, which we abbreviate ST-LP, is robust in the sense it does not require prior statistical knowledge of the channel. The only assumptions are the least-square (LS) estimates have limited spectrum and the channel follows the tapped delay line (TDL) model, which are commonly taken in practice. Since it is desirable slow delay variations operability acceptance, the delay-subspace is tracked by a subspace tracking (ST) algorithm. The ST-LP estimator can be implemented by two filtering structures, which provide a trade-off between accuracy and complexity. Simulation results confirm the superior performance of the ST-LP estimator when compared to methods already reported in the literature.